LiteLLM Server
OpenAI-compatible endpoint powered by 4× GTX 1080 Ti GPUs for no-cost prototyping and course assignments.
MLflow Tracking
Central experiment tracking for coursework and research projects with artifact storage on the lab NAS.
GPU & HPC Access
Internal nodes plus campus clusters cover everything from rapid iteration to large-scale training.
- 8× NVIDIA A30 node for distributed training.
- 2× Titan-class node for low-latency experiments.
- RedHawk, TALON, and OSC allocations for large jobs.
LabelMe Instance
Lightweight annotation environment for bounding boxes and polygons.